Decision tree induction shows potential for the detection of clinical mastitis

C. Kamphuis; H. Mollenhorst; A. Feelders; H. Hogeveen

Decision tree induction shows potential for the detection of clinical mastitis

In: Precision livestock farming '09

Authors:

C. Kamphuis

C. Kamphuis ¹1Department of Farm Animal Health, Faculty of Veterinary Medicine, Utrecht University, Marburglaan 2, 3584 CN Utrecht, the Netherlands; C.Kamphuis@uu.nl

Search for other papers by C. Kamphuis in
Current site
Google Scholar
PubMed

H. Mollenhorst

H. Mollenhorst ¹1Department of Farm Animal Health, Faculty of Veterinary Medicine, Utrecht University, Marburglaan 2, 3584 CN Utrecht, the Netherlands; C.Kamphuis@uu.nl

Search for other papers by H. Mollenhorst in
Current site
Google Scholar
PubMed

A. Feelders

A. Feelders ²Department of Information and Computing Sciences, Utrecht University, Padualaan 14, 3508 TB Utrecht, the Netherlands

Search for other papers by A. Feelders in
Current site
Google Scholar
PubMed

, and

H. Hogeveen

H. Hogeveen ¹1Department of Farm Animal Health, Faculty of Veterinary Medicine, Utrecht University, Marburglaan 2, 3584 CN Utrecht, the Netherlands; C.Kamphuis@uu.nl
³Business Economics Group, Wageningen University, Hollandseweg 1, 6706 KN Wageningen, the Netherlands

Search for other papers by H. Hogeveen in
Current site
Google Scholar
PubMed

Type:: Chapter

Pages:: 297–306

DOI:: https://doi.org/10.3920/9789086866632_037

Purchase instant access (PDF download and unlimited online access):

€42.00€35.00 excl. VAT

Add to Cart

Redeem Access Token

This study explored the potential of decision tree induction to develop models for the detection of clinical mastitis with automatic milking. Sensor data (including electrical conductivity, colour and yield) of over 680,000 quarter milkings were collected from December 2006 till August 2007 at six dairy herds using automatic milking. Farmers recordings of quarter milkings that showed clear signs of mastitis were considered as gold standard positive cases (n=100). Those that were not visually checked, that were checked but scored as being normal, that were outside a 2-week range before or after a gold standard positive case and that were not separated manually or automatically were used as gold standard negative cases, of which a random sample of 3,000 cases were selected for further analyses. Decision trees, with varying confidence factors and cost matrices to study their effect on performance characteristics, were developed with the probability of having clinical mastitis for each quarter milking as output. Using 10-fold cross validation, detection performances (sensitivity, specificity) were computed at a threshold of 0.50 for the probability estimate for clinical mastitis. Receiver-operating characteristic curves were constructed to visualize all potential combinations of sensitivity and specificity. In order to study the performance in the range of SP of interest (SP≥97%) the partial area under the curve was computed. Results show that decision tree induction provides models that are easy to interpret when visualised. Although we used field data, labelled records as negative cases of CM of which we were uncertain whether they were truly negative, used a limited time-window in which a classification was valid, and used a simple algorithm to build DTs, performance is similar to those of models currently used by AM systems. Therefore, we conclude that decision tree induction does show potential for detecting clinical mastitis with automatic milking.

Citation Info

Save
Cite
Email this content

Share link with colleague or librarian

You can email a link to this page to a colleague or librarian:
Email this content
or copy the link directly:

https://brill.edhh.ma/abstract/book/9789086866632/BP000037.xml
The link was not copied. Your current browser may not support copying via this button.

Link copied successfully

Collapse
Expand

Precision livestock farming '09

E-Book ISBN:: 9789086866632

Publisher:: Wageningen Academic

Print Publication Date:: 30 Oct 2009

Subjects
- Life Sciences
  - General
  - Animal & Veterinary

Metrics

	All Time	Past 365 days	Past 30 Days
Abstract Views	65	16	3
Full Text Views	3	0	0
PDF Views & Downloads	2	0	0

African Studies	Education	Media Studies
American Studies	History	Middle East and Islamic Studies
Ancient Near East and Egypt	Human Rights and Humanitarian Law	Musicology
Art History	International Law	Philosophy
Asian Studies	International Relations	Religious Studies
Biblical Studies	Jewish Studies	Slavic and Eurasian Studies
Biology	Languages and Linguistics	Social Sciences
Book History and Cartography	Life Sciences	Theology and World Christianity
Classical Studies	Literature and Cultural Studies

Subjects

Authors

Open Access

Product Information

Company

Contact

Decision tree induction shows potential for the detection of clinical mastitis

Citation Info

Share link with colleague or librarian

Precision livestock farming '09

Table of Contents

Metrics

Metrics